Researchers at the University of Waterloo in Canada have examined the accuracy-efficiency tradeoffs of neural language models (NLMs) specifically applied to mobile devices, and proposed a simple technique to recover some perplexity—a measure of a language model's performance—using a negligible amount of memory.
NLMs are language models based on neural networks through which algorithms can learn the typical distribution of sequences of words and make predictions about the next word in a sentence.
Two existing applications use neural networks to provide next-word prediction, but they often require a lot of power to function, draining the batteries of mobile devices.
The researchers assessed several interference-time pruning techniques on quasi-recurrent neural networks, and they suggest training and storing single-rank weight updates at desired operating points to improve performance using a small amount of memory.
From Tech Xplore
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA
No entries found